Skip to main content
Figure 3 | BMC Bioinformatics

Figure 3

From: HECTAR: A method to predict subcellular targeting in heterokonts

Figure 3

Hierarchical architecture of HECTAR. Five categories of subcellar targeting can be predicted by the HECTAR method: signal peptides, type II signal anchors, chloroplast transit peptides, mitochondrion transit peptides and proteins with no detectable N-terminal target peptide. Each decision module (yellow boxes) runs several selected methods (white boxes) to detect specific target peptides. Selected outputs from these methods are then submitted to a SVM which combines these predictors to determine whether a particular target peptide is present in the sequence being analysed. Protein sequences are first analysed by the "signal peptide/anchor" module where a multi-class SVM determines whether a signal peptide or a type II signal anchor is present. If a signal peptide is detected, this sequence is removed from the N-terminal end of the protein sequence and the modified sequence is analysed by the "chloroplast targeted" module which determines whether the signal peptide is followed by a chloroplast transit peptide. In this module the result of a search for the ASAFAP motif is included in the decision process. If a chloroplast transit peptide is present the protein is classified as chloroplastic, otherwise it is classified as having either a signal peptide or a type II signal anchor. When the "signal peptide/anchor" module did not predict either a signal peptide or a type II signal anchor, the protein sequence is analysed by the "mitochondrion targeted" module to determine whether a potential mitochondrion target peptide is present. If a mitochondrion target peptide is found, the protein is assigned as being targeted to the mitochondrion, otherwise it is classified as having no N-terminal target peptide.

Back to article page